The price of fairness: Experimental evidence on the limits to demand for redistribution

How far are people willing to go to reduce inequality and enact what they perceive as a fair distribution of income? We address this question in South Africa, one of the most unequal countries in the world, where questions of fairness and redistribution are part of everyday considerations. We carry out an experiment where a third-party stakeholder is paired with two "workers" and makes a series of redistributive decisions over their earnings. These redistributive decisions carry varying personal costs to the stakeholder, and the inequality in worker payout is also varied. Additionally, we randomly vary the source of inequality: luck or merit. We show that stakeholders' willingness to redistribute increases with initial pay inequality and decreases with personal cost to redistribute. The source of inequality also makes a difference, with higher redistribution in the luck treatment. The source matters less at higher levels of inequality, suggesting a degree of aversion to extreme disparities, even when payouts are seen as rightfully earned. On the other hand, the effect of stakeholders' personal cost does not interact with the source of inequality, indicating a robust self-interest motivation. The interplay of these effects can result in significantly different levels of post-redistribution inequality. We suggest that substantial redistribution might be acceptable to most as a tool for reducing high inequality - such as that observed in South Africa - especially when income allocation is deemed unfair. However, selfinterest may be a significant limiting factor.